Missing structure fields: - warnings

[jyn514]

Not sure how to reproduce, but my code looks like this:

use std::collections::VecDeque;
use std::fmt;
use thiserror::Error;

use super::{Locatable, Location};

Everything from use std::fmt; until use super::{Locatable, Location}; is highlighted in red with the message from the title. I'm not sure what it's trying to tell me, cargo check thinks it's fine.

[Amanieu]

I'm getting a similar error on intrusive-collections.

Steps to reproduce:

• git clone https://github.com/Amanieu/intrusive-rs.git
• Open in vscode
• Open src/linked_list.rs

It will show the following errors, but only if the linked_list.rs file is open:

{
	"resource": "/home/amanieu/code/intrusive/src/linked_list.rs",
	"owner": "_generated_diagnostic_collection_name_#0",
	"severity": 8,
	"message": "Missing structure fields:\n- link1\n- link2\n",
	"source": "rust-analyzer",
	"startLineNumber": 1366,
	"startColumn": 21,
	"endLineNumber": 1369,
	"endColumn": 10
}
{
	"resource": "/home/amanieu/code/intrusive/src/linked_list.rs",
	"owner": "_generated_diagnostic_collection_name_#0",
	"severity": 8,
	"message": "no such field",
	"source": "rust-analyzer",
	"startLineNumber": 1367,
	"startColumn": 13,
	"endLineNumber": 1367,
	"endColumn": 30
}

linked_list.rs

[kiljacken]

@jyn514 The best way to help is try to remove as much unrelated code and push a test case that reproduces the issue as a seperate repostory, that way we can debug locally :)

@Amanieu We appear to be resolving the wrong Obj in your case, I'll split it off into another issue and investigate further :)

[killercup]

Seeing that issue too right now, will try to make a repro-repo.

FTR, it looks liek this here:

[killercup]

Alright, cargo new ra_missing_struct_fields_wrong_spans and paste this into main.rs:

fn some() {}
fn items() {}
fn here() {}

fn main() {
    let _x = vec![
        Foo {
            a: 42,
        }
    ];
}

pub struct Foo {
    pub a: i32,
    pub b: i32,
}

This highlights these spans on my machine with a build from master right now:

It seems to me like the span offset is releative the somewhere inside of the macro but rendered as absolute in the file? (The highlight on Foo is from rustc, not rust-analyzer.)

[killercup]

I started working on this FYI but haven yet fixed it.